Use this URL to cite or link to this record in EThOS:
Title: The entangled cyberspace : an integrated approach for predicting cyber-attacks
Author: Ikwu, Ruth Eneyi
ISNI:       0000 0004 7972 8343
Awarding Body: Brunel University London
Current Institution: Brunel University
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
Significant studies in cyber defence analysis have predominantly revolved around a single linear analysis of information from a single source of evidence (The Network). These studies were limited in their ability to understand the dynamics of entanglements related to cyber-incidents. This research integrates evidence beyond the network in an attempt to understand and predict phases of the kill-chain across the information space. This research provides a multi-dimensional phased analysis of the traditional kill-chain model using structural vector autoregressive models. In the 'Entangled Cyberspace Framework', each phase of the kill-chain corresponds to a single dimension of the information space based on time observations of certain events. Events are represented as time signals, where each phase is characterised by multiple time signals representing multiple events on that phase. Multiple time signals are analysed using structural models for multiple time series analysis (Vector Auto-Regressive models). At each phase of the kill-chain, we perform a lagged co-integration analysis of events across the information space. This nature of analysis detects hidden entanglements that characterise events in the kill-chain beyond the network. The measured prediction accuracy and error measured at each stage of the experiment represents the usefulness of selected events in characterising the defined stage of the kill-chain. The entangled cyberspace, in theory, is the fusion of three conceptual foundations: a) A multi-dimensional characterisation of cyberspace, b) A sequential phased model for perpetrating cyber-attacks and c) A structural model for integrating and simultaneously analysing multiple sources of evidence. It starts with the characterisation of the information space into different dimensions of interest. The framework goes further to identify evidence sources across these characterised dimensions and integrates them in the analytical context under consideration (e.g. Malware Injection). The concrete findings show that our approach and analytical methodology are capable of detecting entanglements when applied to a set of entangled activities across the information space. The findings also prove that activities beyond the network have significant effects on the nature of the unfolding cyber-attack vector. The predictive features of events across the kill-chain were also presented in this research as opinion and emotion drivers on the social dimension, packet data details and social and cultural events on the economic layer. Finally, co-integration detected between events across and within dimensions of the information space proves the existence of both inter-dimensional and intra-dimensional entanglements that affect the nature of events unfolding during the kill-chain (from the adversary's point of view). The novelty of this research rests in the ability to hop across the information space for detecting evidential clues of activities that are related-to cyber-incidents. This research also expands the standard multi-dimensional information space to include SPEC factors as indicators of cyber-incidents. This research improves the current information security management model, specifically in the monitoring, analysis and detection phases. This research provides a methodology that accommodates a robust evidence base for understanding the attack surface. Practically, this research provides a basis for creating applications and tools for protecting critical national infrastructure by integrating data from social platforms, real-world political, cultural and economic events and the cyber-physical.
Supervisor: Louvieris, P. ; Wang, Z. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Kill-chain ; Spec factors ; Vector-autoregression ; Time series ; Natural language processing